Picture this: the used car market, a classic example often cited in economics to explain the ‘Lemons Theory.’ In this market, good cars (‘peaches’) and bad cars (‘lemons’) are mixed together, and it’s incredibly challenging to distinguish between the two. Sounds a lot like hiring, doesn’t it? But here’s where it deviates—especially in specialized fields like data science. Unlike used cars, in the realm of job markets, the ‘lemons’ and ‘peaches’ have the ability to choose you. Bad candidates often gravitate towards poorly articulated or vaguely worded job descriptions. Why? Because they believe these listings offer an easier route through the convoluted labyrinth that is the hiring process. Conversely, top-notch candidates are quick to steer clear of these dubious job postings. They interpret the poor wording as a red flag, indicative of a work environment that’s equally disorganized and lacking direction.

Let me share a recent experience that serves as a case-in-point. I often consult with various companies on their data science initiatives, and not long ago, I came across a job posting from a company that can only be described as ‘very big and very proud.’ You know, the kind of organization that takes immense pride in its achievements and is a key player in its industry. One of their executives had just returned from a conference, buzzing with the news that several competitors had started leveraging machine learning models to predict everything from resource expenditure to customer behavior—and with remarkable success.
Eager to hop on the bandwagon, this executive returned and gave the directive: ‘Find me an outstanding data scientist to spearhead our yet-to-be-formed data science team.’ And so, the recruiters got to work. With no existing expertise in data science, they resorted to cobbling together requirements from multiple existing job postings. The mindset was: ‘If it’s important to one company, let’s add it. If another company demands it, let’s include that too.’ The result? A Frankenstein’s monster of a job posting that seemed to demand experience dating back to the Big Bang.

To say that this posting attracted zero quality candidates would be an understatement. It was like a neon sign flashing, ‘We have no clue what we’re looking for.’ The requirements were a jumble—calling for a ‘practical researcher,’ who is ‘an expert in everything under the sun.’ The company was on the brink of hiring a complete charlatan before I stepped in. Just in the nick of time, I was able to guide them away from what would have been a disastrous hire.

Let’s play out a nightmare scenario, shall we? Imagine this charlatan takes the reins. Suddenly, they’re in a seat of power, tasked with building a team from scratch, dictating workloads, reviewing methodologies, and essentially shaping the entire trajectory of this fledgling data science department. So how would someone unqualified navigate this maze while safeguarding their own interests?
For starters, they’d hire equally unqualified candidates. After all, a weak team won’t be able to reveal the leader’s own incompetence. Next, they’d opt for projects that don’t promise any immediate returns—long-term endeavors that keep deliverables conveniently distant. This would create a smokescreen, pushing any real evaluation of their capabilities further down the calendar.

And when things inevitably start falling apart? This charlatan would be armed with a litany of believable excuses to shift the blame elsewhere—be it the market conditions, the inadequacies of other departments, or some other scapegoat.

By the time the dust settles and the company realizes they’ve been led astray, we’re talking about half a year gone by, if not more. And what’s left in the wake of this calamity? A data science team that’s about as effective as a screen door on a submarine. Worse still, after this debacle, the company would find itself in a precarious position, ill-equipped to attract the kind of skilled leader they initially sought but failed to understand.
So, how should a company go about this the right way? First and foremost, if you’re venturing into unfamiliar territory, you’ll need a knowledgeable guide—someone who can oversee the hiring process from the get-go. Now, I can already hear the collective gasps at the potential price tag, but let me break down what this actually involves:

1. Understanding Needs: The guide sits down with hiring managers to truly comprehend the needs and goals of the new team.
2. Org Structure: They’ll help define the makeup of the team—how many senior roles are needed, how many junior roles, and so on.
3. Budgeting: Programmers don’t come cheap, so budgeting becomes a critical step. Can the company afford the talent it needs?
4. Crafting Postings: Next up is sculpting job postings that not only make sense but are also enticing to qualified candidates.
5. HR Screening: This guide instructs the HR team on what to look for during initial screenings.
6. Quality Checks: Regular reviews ensure the screening process stays on track.
7. Technical Interviews: The guide also preps technical interviewers, offering advice on dos and don’ts, and even sits in on a few interviews to evaluate the process.
8. Iteration: If the interviews aren’t up to par, it’s back to the drawing board. Yes, this may involve finding a new interviewer and starting over.
9. Direct Involvement: Sometimes, it gets to the point where the guide loses faith in humanity and steps in to conduct the technical interviews themselves.
10. Offer Presentation: Lastly, the guide helps the hiring manager in framing the offer.
11. Onboarding: Oh, and did I mention there’s onboarding after all this?

Soon, I’ll delve into the costs associated with getting this process wrong. Spoiler alert: it’s even more expensive than you think.